Hi Don,
On 4/22/20 5:24 PM, Don Murray via netcdf-java wrote:
Hi All-
I'm looking at storing forecast model data (NMME and GEFS) in netCDF 
format and am wondering if anyone out there is doing that and using 
netcdf-java to access and subset.
Yes. You can take a look at:
https://meteo.unican.es/trac/wiki/udg/ecoms/dataserver/catalog
In particular System4, CFSv2 and glosea5 datasets which corresponds to 
different seasonal forecasts systems.
For GRIB data, netcdf-java/THREDDS uses a 2D time dimension with 
dimensions (reftime, time, ens, lat, lon) where time is stored as
time(reftime, time). I'm looking for examples that use netCDF as the 
storage format, particularly if they are served up via THREDDS.  I'm 
using the ucar.nc2.dt.grid classes to read netCDF files at the present.
In particular we are using netcdf-java for accessing data from client 
side and TDS from server side to aggregate data in common datasets.
In the attached files you can find the opendap info for on of the 
datasets which includes "realization" coordinate and 2D time, among 
level, lat and lon coordinates (level it's a particular case).
https://www.unidata.ucar.edu/software/netcdf-java/v4.3/v4.2/javadocAll/ucar/nc2/ft/grid/Grid.html#makeSubset(ucar.ma2.Range,%20ucar.ma2.Range,%20ucar.ma2.Range,%20ucar.ma2.Range,%20ucar.ma2.Range,%20ucar.ma2.Range)
The client use de makeSubset method from the Grid object which allows to 
subset along the realization coordinate and 2D time:
https://www.unidata.ucar.edu/software/netcdf-java/v4.3/v4.2/javadocAll/ucar/nc2/ft/grid/Grid.html
The dataset are the original GRIB1 files aggregated using netcdf-java 
ncml aggregation and published it by TDS.
This is been used from a different R packages we developed, but in 
particular the central part it's the loadeR which interacts directly 
with the netcdf-java API (by rJava):
https://github.com/SantanderMetGroup/loadeR
You have some academic reference and use cases there.
Please, let me know if you need more info.
Regards
Antonio
--
Antonio S. Cofiño
Dep. de Matemática Aplicada y
        Ciencias de la Computación
Universidad de Cantabria
http://www.meteo.unican.es